Compact M.2 inference accelerator for power-constrained vision devices.

Integrated compute board for building embedded AI applications.

Europa-powered PCIe card for higher-capacity physical and enterprise AI inference.

Expanded-memory M.2 accelerator for local language, vision-language and multicamera models.

Single-Metis PCIe accelerator for vision and local language inference.

Four-Metis PCIe card for scaling vision and generative AI inference.